Transport Phenomena in Metallurgical

Processes (MMC 401)

Dr. Manas Kumar Mondal

Associate Professor,
Department of Metallurgical and Materials Engineering
National Institute of Technology, Durgapur
 Disclaimer
The study materials/presentations are solely meant for academic purposes and they can
be reused, reproduced, modified, and distributed by others for academic purposes only
with proper acknowledgements.

Lecture 19
References:
 D.R. Poirier and G.H. Geiger, “Transport Phenomena in Materials Processing”, (Springer International Publishers. Switzerland,2016 ).
 Julian Szekely and N.J. Thermelis, “Rate Phenomena in Process Metallurgy” ,(John Wiley & Sons Inc (1 November 1971).
 R. Byron Bird, Warren E. Stewart, and Edwin N. Lightfoot, “TRANSPORT PHEOMENA”, (OHN WILEY & SONS, Inc., 2002).
 D. R. Gaskell, “An Introduction to Transport Phenomena in Materials Engineering”, (MOMENTUM PRESS, LLC, NEW JERSEY,
2013).
 S. P. Sukhatme, “A Textbook on Heat Transfer”, Fourth Edition (University Press India Ltd., 2005).

 J. P. Holman, “Heat Transfer”, (The McGraw-Hill Companies, Inc., )

MASS TRANSPORT
Diffusion in Gases

 The initial attempts for predication of diffusion coefficient in gases and mixtures were based on kinetic theory of ideal
gases.
 According to this theory, the diffusion coefficient, DAB of a species A at low gas density in a mixture of gases A and B of
equal molecular weights and sizes is expressed by the reaction

1
D  v .....................(1)
3
AB A

where v is the root mean square velocity of the gas molecules expressed by relation
1

 8 RT 
2

v    ....................(2)
 M N
A 

and A, for component A, is the mean free path of the gas molecules, expressed by the relation

1
  ..................(3)
2n d
A 2
V A

where dA is the diameter of the molecules of gas A, nV the number of molecules per unit volume and MA is the molecular
weight of gas A.
For ideal gas the number of molecule per unit volume can be determined from Avogadro’s number and the ideal gas law
leading to
nN nN NP
n    ................( 4)
V nRT RT
V

P
RT
Substitute equation (4) into equation (3) leads to   .......... .( 5 )
2 NPd
A 2
A

1 1

1  8 RT  2 R T  1
2 2
RT 3 3

Now substitute equation (5) and (2) in equation (1) leads to D       .....(6)
3  M N  2NPd 3 M N Pd
AB 2 3 3 2
A A A  A

Calculation of values of DAB using the above equation yields results for molecules of same molecular weight and diameter
for non-polar gases i.e. for isotopes of non-polar gases, within about 5% of the experimental data. Extension of this
approach to nonpolar gases of different molecular weights and sizes leads to the equation
3 1 3 1

 4   2 2 
2 2
 RT  4 1 k T  4
2 2

D       
B
   ..............(7)
 N   3P d  d     M  3P d  d 
2 2

 2M 2M M
AB
A B
 A

B A B A B

 2 

Where N is the Avogadro’s number, kB is Boltzmann’s constant = 1.38  1023 J/K or 1.38  1016 erg/K, dA and dB are the
diameters of atoms or molecular gases A and B, respectively and MA and MB their respective atomic or molecular weight.
Chapman and Enskog developed more accurate relation for diffusion coefficient by taking into account the interactions
among various molecules. Their equation of gas A in a binary mixture of gases A and B at low density is as follows:
1

2.2646  10   1 
2
1 5

D  T    ...............(8)
C   M M
AB 2
AB D , AB A B 

0
where the concentration C is expressed in g.mol.cm3, temperature T in K, function AB in A and function D,AB is a
dimensionless function. Its values as a function of kT/AB are given in table 1.

Function AB and AB, are called Lennard-Jones parameters for the gas mixture, may be calculated from the standard
values of these for the constituent gases from Table 2 by making use of the following equations

AB = 0.5 (A + B) …………………. (9)

AB = (AB)2 …………………. (10)

Table 1: Transport characteristic function of gases
kT/ or kT/AB D,AB (for mass diffusivity)
0.3 2.662
0.35 2.476
0.4 2.318
0.45 2.184
0.5 2.066
0.55 1.966
0.6 1.877
0.65 1.798
1.00 1.439
2.00 1.075
3.00 0.949
4.00 0.8788
5.00 0.7712
90.00 0.5256
100.00 0.5170
 Table 2: Characteristic of gases
Gas Molecular Lennard-Jones Critical constant
weight parameter
Mi A /k (K) Tc (K) Pc(atm) Vc (cm3/g mole)
Air 28.97 3.617 97.0 132 36.4 86.6
Argon 39.944 3.418 124.0 151 48.0 75.2
Bromine 159.83 4.268 520.0 584.0 102.0 144.0
Carbon 25.01 3.590 110.0 133.0 34.5 93.1
monoxide
Chlorine 70.91 4.115 357.0 417.0 76.1 124.0
Carbon 44.01 3.996 190.0 304.2 72.9 94.0
dioxide
In the case of vapors of metals, the value of A can be taken equal to the atomic diameter of the metal.

Apart from the above, another relation is frequently used for the calculation of diffusion coefficients in a binary as mixture
at low pressure. This relation is based on the concept of corresponding states and is written as
b

a T 
12
 1 1 
D    P P  T     .................(11)
13 5 12
T
P  T T 
1
M M
AB c,A c ,B c ,A c ,B
 
2
c,A c ,B  A B

Where the symbols Pc,A and Tc,A represent the critical pressure and temperature, respectively for gas A and Pc,B and Tc,B
respectively those for gas B, and a and b are constants whose values are given as follows:

(i) For nonpolar gas-pairs

a = 2.745 104 and b = 1.823

(i) For water vapor with a nonpolar gas

a = 3.64 104 and b = 2.334
Values of self-diffusion coefficients at other temperatures and pressure can be calculated at other temperatures and pressure
can be calculated with help of a graphical relationship between PDA/(PDA)0 and Pr.

Figure : Relationship between function (PDA)/(PDA)0 and reduced pressure Pr at different reduced temperatures for gases.

In this figure (PDA)0 refers to the diffusion coefficient at low pressure for the temperature under consideration.

Tr and Pr are the reduced temperature and pressure respectively and are calculated as the ratio of the concerned temperature
and pressure to the critical values of these variables for the gas the under consideration.
Problem: Calculate the self-diffusion coefficient of argon under 0.2 atmosphere pressure and 298 (i) by making use
of kinetic theory of ideal gases and (ii) with the help of Chapman-Enskog equation. The relevant data for argon
0 0

and other constants are as follows: Atomic weight = 40, atomic diameter = 3.82 A ,  = 3.418 A , /k = 124 K, R =
8.314  103 J.kg-mole1 K1, N = 6.02  1023.

Solution:

(i) According to kinetic theory of gas

2  8.314  10  298 
1 1

2 R T  1
2 3 3
1
3 3 3 2

D      
AB
3 M N 3
A
3
 Pd 3   2
A
40  6. 02  10 
3

23 3
0.2  1.01  10 3.82  10
5 10

2

= 6.23  105 m2/s

(ii) For Argon gas D  2.2646  10  2T 

5 2

C  M 
AB 2
AB D , AB A
 Under the assumption of ideal gas law, we can write

C = n/V = P/RT = 0.2/(8.314  298) = 8.18  106 g.mole.cm3

Also kT/ = 298/124 = 2.4 from Table 1, we can write D,AA = 1.012

1
 2.2646  10 5
  2  298  2

D      0.9 = cm2 s1 = 9  105 m2s1

 8.18  10  3.418    40 
AB 6 2

 A comparison of this result with that calculated using the equation based on the kinetic theory shows that the two
differ by about 30%.

 It has also been observed that the calculated value obtained by the second method approximates the experimental
data.

 This anomaly is due to the fact that derivation of the expression based on the kinetic theory of gases does not take
into account the interaction between different gas molecules. Thus, this calculation shows that non-consideration of
neutral interaction between different gas molecules can lead to results which deviate considerably from the
experimental data.
Problem: Calculate the diffusion coefficient of oxygen in air at one atmospheric pressure and 25C using following
data MO2 = 32, TC,O2 = 154.4 K, PC,O = 49.7 atm, TC,N = 126.2K, PC,N2 = 33.5 atm and MN = 28.0
2 2 2
Also calculate the diffusion coefficient of oxygen at 100 atm pressure and temperature remaining unchanged.
b
Solution:
12
a  T   1 1 
DO   P P  T13
Tc , N  5 12
  where a = 2.745 104 and b = 1.823
2
P  Tc ,O Tc , N
2 2
12  c ,O c , N2 2 c , O2 2 M
 O MN 
2

2

b
2.745  10 4  
12
298 5 12  1 1
DO  1  49.7  33.5  154.4  126.2   
13
 
1  154.4  126.2   32 28 
2
2

= 0.207 cm2s1 = 2.07  105 m2s1

P 100 T 298
Further, Pr ,O   and T    1.93
2
Pc ,O 33.5
2
r ,O2
Tc ,O 154.4 2

Now from plot between function PDA/(PDA)0 and reduced pressure Pr at different reduced temperature for gases. We get
{PD/(PD)0} = 0.9

0 .9
Thus, at 100 atm pressure D O2 ,100 atm
 1 0.207  10 4   1.86  10 7 m2s1
100
 Further, Jeans, Chapman and Sutherland derived expression for diffusion coefficient for monatomic gases using the
kinetic theory. Gilliland used the same form of equation and determined the constants empirically based on published
experimental data and proposed the following equation for diffusion coefficient in a binary gas mixture
3
T 2
1 1
D  435.7  ........................(12)
 1
P V A 3  VB 3
1

2
MA MB

Where D is in cm2/s, T is in K, P is total pressure in N/m2 and VA and VB are atomic volume of species as shown in
Table 3. MA and MB are the molecular weights. For compounds atomic volume can be calculated by using Kopp’s law
of additive volumes.
Table 3: Atomic Volume of species

Air Bromine Carbon Chlorine as Chlorine as Fluorine Hydrogen Nitrogen Oxygen

in R-Cl in R-CHCl-
R
29.9 27 14.8 21.6 24.6 8.7 14.3 15.6 7.4

Equation (11) should be used where experimental data is not available. It shows that D  T3/2.
Problem: Calculate the diffusion coefficient for CO2 – air at 1 atmospheric and 30C temperature.

Solution: VCO2 = 29.6 and Vair = 29.9 and molecular weight of CO2 and air MCO2 = 44 and Mair = 28.9.

435.7  303 2
3 3
T 2
1 1 1 1
D  435.7   
 
2 1
44 28.9 = 0.1 cm s
2 2
MA MB
1.01325  10  34  29.9 
1 1 1
P VA 3  VB 3 5
1
3 3

 

Diffusion in Liquids
Theories on prediction of diffusion coefficients in liquids are based either on the hydrodynamics of fluid flow or on the
absolute reaction theory. There are two basic approaches in theories based on hydrodynamics. The first approach assumes
no slip between the diffusing particle and the surrounding fluid. According to this approach, the diffusion coefficient DAB of
component A in solvent B at temperature T is calculated from the equation

kT
DAB  ......................(1)
6  BRA

Where k is the Boltzmann constant, B the viscosity of solvent B and RA the radius of the molecules of component A. The
above equation is known as Stokes-Einstein equation.
In the second hydrodynamics approach, we also take into account the slip between the particles and the fluid and hence
arrive at following equation
kT
DAB  ................( 2)
4 B RA

Where the symbols have the same meaning as in equation (1).

Both of these expressions give approximate values of the diffusion coefficient. In order to obtain a more accurate result
within  10 percent of the experimental values, the following empirical equation was suggested by Wilke and Chang as an
extension of the Stokes-Einstein equation:

 M B  2T
1

DAB  7.4  10 8
 B
cm 2 / s...................(3)
VA0.6

Where B is called the association parameter for the solvent B and has a value of 2.6 for water,
VA the molar volume of component A in cm3g-mole1,
 is the viscosity of the solution in centipoises
MB the molecular weight of solute component B in solvent A.
Problem: Calculate the diffusion coefficient at 25C of a component A having MA = 178 and ρA = 1280 kg/m3 in water.
Assume viscosity of water as 1 centipoise H2O = 2.6, MH2O = 18.

Solution:

Specific volume VA of a component A is V  178  0.139 m3 kg.mole1 = 139 cm3 g.mole1.

A
1280

 M B  2T 2.6  18 2  298  7.8  10 8 cm 2 s 1

1 1

Now according to modify Stokes-Einstein equation D  7.4  10 8

 B
 7 .4  10 8

VA0.6 1 139
AB 0 .6

Diffusion in solids

 Fick’s first law of diffusion is applied to experimental determination of diffusion coefficients of solid components having
concentration gradient in the direction of movement in a homogenous system composed of solids and called a solid
solution.
 The actual mechanism of diffusion or mode of transfer in solids also be known or studied. From the view point of study
of the mode or mechanism of diffusion, the solids can be classified into the following three categories:

 Interstitial solutions
 Substitutional Solutions
 Ionic Compounds
Diffusion in Interstitial solutions
 In such systems, the diffusing component occupies interstitial positions.
 Diffusion of the component proceeds by a series of jumps of the atoms of this component from the original interstitial
sites occupied by them to one of the neighboring sites.
 For example, carbon occupies octahedral interstitial position or site in -iron and its diffusion proceeds by the jump of
carbon atoms from one such site to the other.

 Diffusion coefficient for such a component i can be calculated using the equation

 G 
Di  a 2 p exp  .............(1)
 RT 
H 
Which can also be written D  a p exp
2 S   H 
 exp  .............( 2) or D  D exp 
i 0 .............(3)
 RT 
i
 RT   RT 
 S 
where H = activation energy for diffusion and D0  a 2 p exp .............(3)
 RT 

where,  is a geometric factor, the value of which depends upon the of crystal and is equal to 1/2 for FCC and 1/24 for BCC;
a is the lattice parameter for the solvent;
p is the number neighboring interstitial sites to any specific site from which the component is diffusing;
 is the vibrational frequency of an atom in an interstitial site;
and G is the free energy change for an atom to jump from an interstitial site to the neighboring one.
 For experimental determination of diffusion coefficients of interstitial components, equation (1) rewritten in the
following alternative form:
a 2
D .................( 4)


where  is the mean time of stay of the solute atom in an interstitial site.

 Diffusion coefficients of interstitial components are usually determined by the internal friction method which yields the
value of relaxation time  for inelastic strain.

 This relaxation time is related to  for a BCC lattice by the equation

3
    ....................(5)
2
Thus, equation (4) assumes the form
1 a2
D a2  .................(6)
24  35 

This equation is commonly used for the calculation of diffusion coefficients of interstitial solutes
Problem: Calculate the diffusion coefficient of carbon in -iron at 500C from the following data:

Lattice constant of iron = 2.85 A
Inelastic strain relaxation time () = 4.7  1010 s
Also, calculate (i) the activation energy H, and frequency factor D0 for diffusion if the value of  changes to 4.24  108
s at 300C and (ii) the mean time of stay of a carbon atom at an interstitial sites. Neglect the effect of change in lattice
parameter caused by temperature variation.

Solution:
According to equation (6), value of diffusion coefficients, D773 and D573 at 500C and 300C are respectively calculated as

D773 
a2

2.85  10 8 
2

 4.8  10 8 cm 2 s 1  4.8  10 12 m 2 s 1

36  36  4.7  10 10

and
D573 
a2

2.85  10 8 
2

 5.3  10 14 m 2 s 1
36  36  4.28  10 10

H
Further, according to equation (3), we have ln D  ln D  0
RT
H
Substituting the relevant values of D773 and D573, we get ln 4.8  10   ln D  12
 26.06  ln D0  1.56  10 4  H
8.314  773
0
 H
and similarly ln 5.3  10   ln D 
12
 30.57  ln D0  2.10  10 4  H
8.314  573
0

which on simplification yield the required values as

H = 83.5 kJ lnD0 =  13.119 D0 = 2.01  106 m2s1

Further, using equation (5) for mean time, its value at the two temperatures can be calculated as

3
    ....................(5)
2

At 773 K 3
 773   4.7  1010  7.05  1010 s
2

At 573 K 3
 773   4.28  10 8  6.42  10 10 s
2